[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Adding lzma to dpkg's Pre-Depends



On Fri, Apr 04, 2008 at 03:20:10PM +0200, Julian Andres Klode wrote:
[...]
> I strongly recommend not to use lzma on embedded architectures.

I suppose that depends on how it's used, but Wikipedia's article on
the algorithm even goes so far as to state:

   "Small code size and relatively low memory overhead, particularly
   with smaller dictionary lengths, make the LZMA decompression
   algorithm well-suited to embedded applications."

http://en.wikipedia.org/wiki/Lempel-Ziv-Markov_algorithm

And yes, I understand Wikipedia is a far from reputable source, but
correlate with 7-zip's 7z format page which says:

   "The LZMA compression algorithm is very suitable for embedded
   applications."

http://www.7-zip.org/7z.html

And on their SDK page:

   "LZMA provides a high compression ratio and very fast
   decompression, so it is very suitable for embedded applications."

http://www.7-zip.org/sdk.html

Again, objective statement or marketing zealotry? As an embedded
platform use-case example, here's a page talking about (among other
things) work in progress back in 2006 (no idea if it came to
fruition) using LZMA for JFFS2 filesystem compression on OpenWRT:

http://bitsum.com/openwiking/owbase/ow.asp?Embedded_Linux_Compression_Optimizations

I understand that decreasing the dictionary length has implications
on the compression ratio, but it would be interesting to see some
numbers indicating to what degree this is the case (and roughly
where the break-even is against whatever gzip compression level the
autobuilders currently use for packages). Perhaps it's worth
considering adjusting this parameter downward for smaller arch all
packages and binary packages built for typically lower-resource
architectures (m68k, mipsel, arm, et cetera), if there's still an
all-around performance gain to be had over gzip.
-- 
{ IRL(Jeremy_Stanley); PGP(9E8DFF2E4F5995F8FEADDC5829ABF7441FB84657);
SMTP(fungi@yuggoth.org); IRC(fungi@irc.yuggoth.org#ccl); ICQ(114362511);
AIM(dreadazathoth); YAHOO(crawlingchaoslabs); FINGER(fungi@yuggoth.org);
MUD(fungi@katarsis.mudpy.org:6669); WWW(http://fungi.yuggoth.org/); }


Reply to: